Multidimensional Fechnerian scaling: Perceptual separability. (Q1867364)

In multidimensional scaling one starts with a given set of {psychometric functions}, describing the probability with which a pair of stimuli (in euclidean space) are distinguished, and derives the definition of a {distance} between stimuli reflecting perception differences. Fechnerian scaling imposes regularity axioms on the psychometric functions, restricting their behaviour as the metric distance between stimuli varies, and arrives at a well defined concept of distance between stimuli based on the above functions. The author proposes here a coordinatewise separability axiom on the psychometric functions expressed directly, unlike in previous work, in terms of the stimuli dimensions. This, coupled with Fechnerian axioms, allows to express the Fechnerian distance as a Minkowski power metric on (continuously transformed) stimulus coordinates.